4 comments

[ 2.6 ms ] story [ 16.4 ms ] thread
Hey everyone!

Website: https://www.rosieos.com Loom demo: https://www.loom.com/share/fed7d30acb7940708db8e613aed29279

I'm Aman, and I'm creating Rosie (www.rosieos.com). This product is only a few weeks old, but it's been super helpful so far. I built Rosie because trying to learn and debug via StackOverflow is a complete pain and I spend probably 50% of my time flailing around getting nowhere. So I built Rosie.

Rosie is an AI that answers StackOverflow questions. It is currently focused on the React ecosystem, and answers a couple of different types of questions:

- error messages - how do I do something? - how does X work? - opinions (ex. why use TS vs JS)

Why is Rosie better than StackOverflow?

1. Responses are instant. No crawling through 20 posts or waiting hours for a reply 2. No fear of judgement or toxicity. 3. The more detail you add, the better Rosie gets. Dump all the code you want

In the future, Rosie will integrate with your repositories so it has context on your existing libraries, packages and code-writing style. My goal is to make Rosie more effective than bugging the local staff engineer.

Rosie's powered by OpenAI Codex, I'm happy to explain details in the comments. The product is currently in a closed beta, which is why this isn't a Show HN. Please email me at aman @ rosieos.com if you'd like to get on early!

Hope you guys like it, I'm excited to share more details in the future :)

How do you handle accountability? In other words, how can I trust that the response Rosie gives me is actually helpful and not actually a subtle anti-pattern?

Bonus question: How did you choose the price point? Competitors have priced Codex accessibility at around $10-15 per month. What's the value-add that makes Rosie worth $35 monthly?

I don't see the point of using this when you can just use a search engine to look up the same information for free.

I think it would be more interesting too hook it up to stack overflow and actually answer people's questions.

I thought that was what it was! Then - oh I have to pay?